Found in

• Fx 107.0a1

Affected versions

• Fx 107.0a1
• Fx 106.0b5
• Fx 105.0.1

Affected platforms

• Windows 10
• Ubuntu 20
• macOS

Steps to reproduce

1. Launch Firefox.
2. Open a Private Browsing window.
3. Click the websearch bar type a word and hit enter.
4. Go to about:telemetry -> Raw JSON -> Raw Data and check for the search_counts under keyed.histogram

Expected result

• There is an 'urlbar-handoff' telemetry ping recorded.

Actual result

• Only a 'urlbar' telemetry ping is recorded.

Regression range

• This is not a recent regression. Will look into a regression range ASAP.

AFAIK we've never recorded hand-off telemetry for the private window - the private window has always had the hand-off mode, whereas it was something that was introduced later on for the normal mode (and hence we wanted telemetry for it).

I'm not sure if we need to record this in private mode. We'll probably need to ask product what's required.

Mozregression results:
Bug 1766887 - Test that newtab search issuances appear as Glean events r=Standard8

And slightly reconfigure the SERP test for brevity, and to ensure the newtab
visit id remains constant throughout.

Oddly the search engine telemetryId is slightly different in
SearchSERPTelemetry than in BrowserSearchTelemetry. Might just be a test quirk.

Differential Revision: https://phabricator.services.mozilla.com/D149536

:chutten, since you are the author of the regressor, bug 1766887, could you take a look?

For more information, please visit auto_nag documentation.

There is no code in D149536 that changes how legacy telemetry is recorded, so I'm unsure how that could be the regressor. There are other patches in that stack, though...

Wait, :Standard8, does what you're saying mean that this definitely isn't a regression?

Set release status flags based on info from the regressing bug 1766887

(In reply to Chris H-C :chutten from comment #4)

There is no code in D149536 that changes how legacy telemetry is recorded, so I'm unsure how that could be the regressor. There are other patches in that stack, though...

Wait, :Standard8, does what you're saying mean that this definitely isn't a regression?

I was wrong. We did record this previously.

Having looked into this, I think the regression range is correct - In particular I think this change is causing the issue.

The private browsing window calls into that from a couple of places.

So I think we probably need a second attribute, or something that we use to know that we're in handoff mode, but don't record as a new tab session id.

Oh well, poop. Sorry about this. Lemme get this fixed.

https://hg.mozilla.org/mozilla-central/rev/98d851e31e64

Comment on attachment 9308566 [details]
Bug 1792815 - Reintroduce handoff reporting for pbm-initiates searches r?standard8!

Beta/Release Uplift Approval Request

• User impact if declined: No user impact. Business impact in our ability to accurately count searches from private browsing windows
• Is this code covered by automated tests?: Yes
• Has the fix been verified in Nightly?: Yes
• Needs manual test from QE?: No
• If yes, steps to reproduce:
• List of other uplifts needed: None
• Risk to taking this patch: Low
• Why is the change risky/not risky? (and alternatives if risky): Instrumentation-only change, (manually) verified in nightly, covered by automated tests (though with an intermittent)
• String changes made/needed:
• Is Android affected?: Unknown

Comment on attachment 9308566 [details]
Bug 1792815 - Reintroduce handoff reporting for pbm-initiates searches r?standard8!

Approved for 109.0b6.

https://hg.mozilla.org/releases/mozilla-beta/rev/6f95380a08ca

The issue is verified fixed in the latest Fx 112.0a1 and Fx 111.0b3 on WIndows 10 and Ubuntu 22.04. The handoff telemetry is correctly recorded in private browsing mode with DDG, Google and Bing.